The broad aim of this work is to understand the molecular forces that determine the interaction of proteins with ligands. The goal is the development of the theory needed for molecular design. This project focuses on computer-assisted design of ligands for macromolecular receptors of known structure using a program called DOCK. The current DOCK program creates negative images of receptor surfaces and positive images of ligands. It then explores in a systematic way the six degrees of freedom involved in fitting together two rigid bodies. It does this operation rapidly enough that we can use large structural databases - the Cambridge Structural Database, the Fine Chemicals Directory, and the Chemical Abstracts Registry - as a source of molecular templates for ligand design. In the coming grant period we suggest ways to make very substantial improvements in the performance and speed of the DOCK program, and we suggest direct approaches to the de novo design and optimization of active ligands. The specific proposals for DOCK modifications are: to develop conformational searching for ligand and receptor; to introduce more quantitative scoring by accounting for solvation effects and differential scoring; and to speed up the search process through shape cluster analysis and code improvements. For the ligand design section, we propose to: automate ligand modification procedures; continue development of a de novo ligand construction program called BUILDER; explore "positive" imaging for mimetic design; incorporate pharmacological properties into our databases. The DOCK approach to ligand discovery and design has had some important initial successes in the development of novel lead compounds as inhibitors of the HIV protease; the interruption of the CD4-gp 120 recognition event that precedes HIV infection; and inhibition of influenza virus, among others. DOCK is an important tool in structure-based design and has helped to speed up the drug discovery process.